BioEd Online

When it comes to deadly protein clusters in the brain, size matters.

The human equivalent of mad cow disease, variant Creutzfeldt-Jakob disease (vCJD) is thought to be caused by misshapen proteins, known as prions, that infect the brain. Research now shows that the most infectious strings of prions are of a middling length; clumps that are longer or shorter are less problematic.

The findings, reported in this week's Nature¹, could convince medical experts to rethink how they plan to treat illnesses such as vCJD, as well as Alzheimer's and Parkinson's.

Model hamsters

Researchers have often debated whether longer or shorter chains of prions are more problematic. The molecules seem to multiply by converting the normal proteins that they touch to an irregular form. Long ones form visible tangles in the brain, but short ones might be more capable of spreading the infection.

Jay Silveira and his colleagues at the Rocky Mountain Laboratories in Hamilton, Montana, obtained misshapen prion proteins from hamsters, broke them up using a detergent, and sorted them according to size. They then injected strings of known length into other hamsters.

One group of four hamsters received protein chains that were 21 prions long, and these animals all succumbed to the disease after about 90 days. Hamsters receiving a solution with protein chains more than 300 prions long died after 90 days only if the solution of prions was 70 times more concentrated, making these longer chains some 70 times less dangerous.

Hamsters given six-unit chains stayed healthy for an additional month, and the concentration needed to cause illness indicated that such short clusters are some 600 times less infectious than the 21-prion chains. Very few of the animals receiving chains of five or fewer prions seemed to get sick at all, living a normal hamster lifespan of two years, says Silveira.

Block and break

Several of the experimental treatments being used on prion diseases such as vCJD aim to block the formation or spread of misshapen proteins. In animal models, many of the treatments seem to work better before the onset of symptoms.

Silveira notes that breaking up prion clusters could be another method, if approached cautiously. "If you could break them up small enough, you could potentially remove their pathogenicity," he says. "The problem would be getting to those small ones without going through the intermediate sizes."

The finding also has implications for treatments of Alzheimer's and Parkinson's disease. In both of these illnesses, long protein threads form in the brain. Silveira and colleagues caution against therapies for these diseases that might fragment these molecular chains into more problematic, moderate lengths.